Comparative machinability and neural network-driven optimization of AISI D2 steel using chamfered and conventional tooling

Dry turning of AISI D2 steel requires a balance between productivity, surface integrity, thermal loading, and energy demand. This study compares the machinability of a chamfered Xcel insert and a conventional carbide insert during dry turning of AISI D2 steel using a 2 4 full-factorial design with cutting speed (V CS : 100–150 m/min), feed rate (F R : 0.2–0.3 mm/rev), and depth of cut (D OC : 0.5–1.0 mm). Volumetric material removal rate (VMR), microhardness (MH), turning-zone temperature (TTZ), and power consumption (PC) were measured and analysed using ANOVA, ANN modelling, and NSGA-II optimization. The results showed that the Xcel insert consistently outperformed the conventional carbide insert within the investigated range. The highest measured VMR and MH were 1341.57 cm 3 and 177.78 HV, respectively, whereas the lowest TTZ and PC were 272 °C and 420.28 W under conservative cutting conditions. ANOVA showed that F R was the dominant factor for VMR (27.69%), while V CS dominated MH (74.51%) and TTZ (49.36%); D OC was the strongest contributor to PC (29.55%). Artificial neural network (ANN) models were used as local-response surrogates, and Non-dominated sorting genetic algorithm (NSGA-II) identified an optimum at 149.84 m/min, 0.3 mm/rev, 0.997 mm, and Xcel insert. The study shows that the Xcel geometry is beneficial for increasing productivity in dry turning of AISI D2 without a disproportionate increase in thermal and energy penalties.